AimsThe Selvester scoring system has been derived from ECG parameters for estimating infarct size. However, there is still a lack of evidence for Selvester score as an alternative to cardiac magnetic resonance (CMR) myocardial injury makers for risk stratification and prediction of left ventricular function (LVF) recovery among patients with ST‐segment elevation myocardial infarction (STEMI).Methods and resultsThis multicentre observational study enrolled 328 STEMI patients (88.4% men, 57.3 ± 10.6 years of age) undergoing CMR examination 1 week post‐reperfusion therapy. Patients with baseline left ventricular ejection fraction (LVEF) < 50% underwent a follow‐up CMR 6 months later, categorized into baseline normal LVF (ejection fraction [EF] ≥ 50% at baseline, n = 155); recovered LVF (EF < 50% at baseline and ≥50% after 6 months, n = 69); and reduced LVF (EF < 50% at baseline and after 6 months, n = 104). The median follow‐up was 4 (3–4) years for all patients, with 61 patients experiencing major adverse cardiovascular event (MACEs). Patients with reduced LVF had a higher risk of MACEs than those with baseline normal LVF (P = 0.01), while the recovered LVF group had no significant difference (P > 0.05). A Selvester score >10 doubled the risk of MACEs in patients with systolic dysfunction (1.91 [1.02 to 3.58], P = 0.04). Additionally, Selvester score, baseline LVEF, transmural infarction, and peak CK‐MB were independent predictors of recovered LVF, with Selvester score providing incremental predictive value to peak CK‐MB in predicting recovered LVF (∆AUC = 0.07, P < 0.05).ConclusionsThe Selvester score improves risk stratification among STEMI patients beyond LVEF and provide independent and incremental information to clinical parameters in predicting recovered LVF.